Bot Basics, Chatbot Development

Chatbots: What Do People Want?

If you’re a business owner, support agent, sales agent, or even a marketing agent, it’s most likely that you understand the potential chatbots bring to a business strategy that prioritizes the customer experience.

Chatbots are becoming more helpful as AI improves and automation gets more powerful. But you still need to get customers to use them.

If you’re a business leader, support agent, salesperson, or marketer, you’ve seen how chatbots enable business strategies truly focused on the customer. However, this doesn’t mean all customers know how valuable chatbots can be.

For example, I’m sure you’ve never heard someone say, “I wish I could use a chatbot to schedule an appointment with my dentist.” In real life, chatbots aren’t people’s first idea of a solution. Regardless, conversational AI tech is getting really good. I believe it should be used more often in the right situations.

So how do you know the right situation?

What Do Customers Really Want In Chatbots?

People want their problems solved or to never have existed in the first place.

As a UX consultant, former AI startup founder, and organizer of a 1,000+ person AI group, I’ve spent years teaching business leaders, product people, and other designers how to start with humans. How to start with a real user problem.

The issue with chatbots—and websites, mobile apps, and startups—is that most are solutions looking for a problem.

So why did their bot fail? Did it have too many steps? Was it not funny enough?

No. They started with a solution and they ended up with something nobody wanted.

Their real problem was that users didn’t know their private data was being stored and would prefer if the company never kept it in the first place.

How can your chatbot be more people-centered?

Chatbots are based on conversations. Conversations are the easiest way for us to exchange information (geeks, see “cognitive ergonomics”). In this way, chatbots are already more human-centered than many other technologies.

This explains why they outperform lengthy contact forms and gather better info than apathetic insurance claim wizards.

But, as you know, technology isn’t enough by itself. We need to make something people want.

How? It’s simple.

First, include your users. Have meaningful conversations with different personas. Invite them to co-design the future with you. Sit in on customer service calls and read support requests. Walk a mile in their shoes.

Ask “why” a few times until you learn what they’re really after. In my 1000s of user research hours, I’ve learned almost everyone wants to…

Be happier

Save time

Save money

Feel safe

Help their families and communities

Now, think, “How can I give them these things without getting in the way?”

Here are some top tips:

Get out of the way
What’s your user’s current journey? How can you remove steps instead of adding them? Challenge yourself to find ways to embrace their typical processes.

Don’t handle problems; prevent them
Does this issue need to exist at all? How can you proactively solve it before it needs fixing?

Remove friction
How can you make things faster, easier, and less frustrating? Quicker is usually better.

Be a great friend
Do awesome things users didn’t even ask for. Think, “How would I solve this for a close friend?”

When might you use a chatbot?

Sometimes, a bot is the answer.

I recently talked with Shane Mac, SVP at Conversocial, about what bots are best at. He's helped brands like Sephora, 1-800-FLOWERS, and Fandango creates some of the most successful bots ever made.

Shane mentioned 2 situations where bots might be a great choice.

1. If you want insightful conversations with every user

Messaging is the most scalable way to actually hear what your users have to say. Websites only give you heat maps and click charts, while messaging gives conversational data full of hints of what someone wants. And in Shane’s words: “Repeated hints are patterns. Repeated patterns are preferences. Acknowledging customer preferences let’s you make something they’ll love.”

2. If it’s faster or easier

Channels like Facebook Messenger, SMS, and WhatsApp are special because you can use an app without downloading anything or making a new account. In situations, a bot can handle from start to finish (e.g., booking an appointment, ordering food), this can save people a ton of time.

A lot of Shane’s success with bots comes from using context, preferences, and previous interactions to predict what users want to do next. He ended our talk with, “the future isn’t about automation, but anticipation.”

Be bold

Thinking beyond proven solutions is scary. Employees keep their jobs when they play it safe and get predictable results. But the digital landscape is screaming past everyone in some way, so playing it safe means sure failure.

What if, instead of banning “crazy” ideas from the conference room, we embraced them and searched them for merit?

How might you improve customer trust if you did automatically delete private data? What if it wasn’t “impossible” to save new employees weeks by automating the onboarding process? What if you did invest in a product recommendation algorithm instead of adding more store filters? What if you could get better data with an IoT sensor instead of asking for user input?

All life-changing products spawn from many uncomfortable conversations with the right decision-makers at the right time.

I think it’s time to believe in a better way, speak up, and get excited about a world where choosing a technology happens after answering the question, “What do people really want?”